Instantaneous Current Predcton for Naval Operatons Peter C. Chu Naval Ocean Analyss and Predcton Laboratory, Department of Oceanography Naval Postgraduate School, Monterey, CA 93943 Albert E. Armstrong Naval Ocean Analyss and Predcton Laboratory, Department of Oceanography Naval Postgraduate School, Monterey, CA 93943 Abstract Naval operatons depend hghly upon envronmental condtons that can ether adversely affect successful completon or hnder the safety of personnel. Each warfare communty has defned envronmental thresholds and operatng lmts that restrct the executon of any ntended maneuver. As the warfare envronment contnues to shft from the open ocean to the lttoral, predcton of the shallow water envronment s an urgent need n order to support these operatons. The value-aded of usng a hydrodynamc model (WQMAP for the msson plannng of the naval operatons n San Dego Bay s demonstrated n ths study. A new model verfcaton procedure (.e., compatblty verfcaton s proposed for the tdal domnated lttoral basn predcton.. INTRODUCTION Executon of any naval operaton can be hndered by numerous marne envronmental factors. Proper plannng for these operatons ncludes forecastng the envronment of the area of operatons. Each warfare area has defned envronmental thresholds and operatng lmts that restrct the executon of any ntended maneuver. Amphbous assault vehcles, mechanzed landng craft, utlty landng craft, AN/SLQ-48 mne neutralzaton vehcle aboard Avenger class mne countermeasures shps and Osprey class coastal mne hunter, and dver operatons all have varous ocean current thresholds. Lttoral warfare oceanography s to study the measure of effectveness of envronmental nformaton, to dentfy the envronmental effects on lttoral warfare (mne warfare, specal warfare, homeland securty,, to optmze war-fghtng requrements usng quanttatve nformaton of the envronment, and to make an envronmental plannng ad for the war fghter (Armstrong, 4. Mne warfare (MIW operatons such as mne huntng and clearance are conducted wth the condton that the nstantaneous current speed s less than a threshold (c, V c. If the current speed s larger than c, the operaton should be stopped. Snce the tdal currents domnate the lttoral zone, the Natonal Oceanc and Atmospherc Admnstraton (NOAA tde table that provdes maxmum ebb and flood current speeds s used to schedule the operatonal perod. For example, A MIW operaton was conducted wthn San Dego Bay (Fg. on January 4. The scenaro s that just a day before a freghter extng the bay was seen dschargng several large objects n the vcnty of Staton-4 (nsde San Dego Bay as t steams outbound for sea. A mne countermeasures shp should be dspatched to locate and dentfy the mne lke objects utlzng ts AN/SQQ-3 varable depth mne huntng sonar along wth the AN/SLQ-48 mne neutralzaton vehcle. The current thresholds for these assets are operatonally senstve, so a fcttous and current lmt of m s - (.e., knots wll be used as an operatonal current threshold n ths example.
99-996 observed data []. Dscrepances between model predcton and feld data n both model calbraton and verfcaton are on the order of the magntude of uncertantes n the feld data []. Fg.. San Dego Bay geography and several locatons: Presently, the tdal current obtaned from the NOAA tde table s the only source of envronmental nformaton for the Navy s MIW unts to schedule the mne countermeasure operatons. However, the NOAA tde table data are usually calculated at partcular locatons. For example, n San Dego Bay, the NOAA tdal current data are only avalable n the mouth of the Bay (.e., Staton- wth four data ponts per day: two maxmum ebb speeds and two maxmum flood speeds. Usually, the dspatcher schedules the mne countermeasure shp operatons at Staton-4 s on the base of the NOAA tde table data for Staton- and the lnear nterpolaton of four tmes daly data nto hgh resoluton tme seres. For the tdal current data shown n NOAA tdal table, the mne countermeasure can operate wthn the threshold ( m s - all mornng untl :. The mne huntng would be able to resume agan around 3:45 wth no other restrctons for the rest of the day (Fg.. Questons arse: Can the NOAA tdal current data for Staton- be used for Staton-4? Can the nterpolaton be used to get hgh temporal resoluton data from the four tmes daly NOAA tdal current data? Assessment of exstng technology s essental to answer these questons. Recent advances n numercal modelng and computer hardware have made robust and effcent numercal models avalable. These models can provde suffcent spatal detals to smulate tdal hydrodynamc processes n bays and estuares. A -m resoluton (3,8 grd nodes, depth averaged, numercal hydrodynamc model was mplemented for San Dego Bay to descrbe essental hydrodynamc processes n the bay. Ths two dmensonal model was calbrated usng the 983 observed data and verfed usng Fg.. Msson plannng for mne countermeasure and mne huntng nsde San Dego Bay usng NOAA tde table maxmum ebb and flood current veloctes at Staton-. In ths study, a D hydrodynamc model smlar to [] wth curve-lnear coordnates and graphc use nterface s used to predct nstantaneous current velocty n San Dego Bay wth hgh temporal and spatal resolutons usng NOAA tde table data at Staton-. The model predcted current velocty at Staton-4 (or other locaton nsde San Dego Bay can be used for schedulng the mne countermeasure shp or any other naval operatons. The purpose of ths study s to show the wde applcaton of the lttoral ocean models to the Navy operatons. The rest of ths paper s outlned as follows. Secton descrbes general oceanographc condtons of San Dego Bay. Sectons 3 presents a hgh-resoluton hydrodynamc model (WQMAP wth boundary fttng curve-lnear coordnates and graphc user nterface. Sectons 4-6 show the model mplementaton, tdal forcng, and compatblty verfcaton. Sectons 7-8 llustrate the applcaton of the model predcton to the msson plannng of the naval operatons. Secton 9 presents the conclusons.. SAN DIEGO BAY San Dego Bay s located n southern Calforna near the Mexcan border and s the home port to a large fracton of the U. S. Navy s actve fleet. The bay has been extensvely engneered to accommodate shppng actvtes wth a deep navgaton channel n the northern part of the bay whle the south part s very shallow. About 9% of all avalable marsh lands and 5% of all avalable nter-tdal lands have been reclamed, and dredgng actvtes wthn the
bay have been equally extensve (Peelng, 975. The San Dego Rver was dverted by the U. S. Army Corps of Engneers n 875 and no longer emptes nto San Dego Bay. San Dego Bay s a small crescent shape bay, 43 km n area at mean lower low water (MLLW, wth an average depth of 6.5 m measured from the mean sea level. Maxmum tdal range n the sprng tdes exceeds m, whereas the mean tdal range s about.85 m. Tdal current speeds range between.3-.5 m s - near the nlet and.-. m s - n the southern regon of the bay. Freshwater nflow to the bay s low and only occurs durng nfrequent wnter storms. Westerly afternoon wnds occur throughout the year and are about 5 m s - ; evenng and early mornng easterly wnds occur prmarly n wnter and are less than 5 m s -. Snce freshwater flow n the bay s low as well as wnd magntude, currents are predomnately produced by tdes. Except durng rare occasons, San Dego Bay can be treated as a vertcally well-mxed estuary (Wang et al., 998. 3. HYDRODYNAMIC MODEL The numercal hydrodynamc model mplemented for San Dego Bay s a depth-averaged, boundary ftted tdal and resdual crculaton model known as WQMAP ([4], [5]. The numercal technques ncorporated n the model are well documented, thus only a summary of the model characterstcs s presented. WQMAP s an ntegrated hydrodynamc and water qualty modelng system desgned for use wthn coastal and fresh water envronments. Ths commercal off-the-shelf program was developed by Appled Scence Assocates, Inc. out of Narragansett, Rhode Island. WQMAP conssts of three basc components: a boundary-ftted coordnate grd creaton module, a three-dmensonal hydrodynamcs model, and a water qualty or pollutant transport model. These models are executed on a boundary ftted grd system. They can also be operated on any orthogonal curvlnear grd or a rectangular grd, whch are specal cases of the boundary ftted grd. The model s confgured to run n a vertcally averaged (barotropc mode or as a fully three-dmensonal (baroclnc mode. Several assumptons are made n the model formulaton, ncludng the hydrostatc (shallow water approxmaton, the Boussnesq approxmaton, and ncompressblty. In ths study, the D verson s used. Most strkng feature of WQMAP s ts hybrd orthogonal curvlnear-terran followng coordnate system. Let ( ϕ, λ,z be the lattude, longtude, and heght, and (ξ, ησ, be a hybrd coordnate system wth a generalzed orthogonal curvlnear coordnate system ( ξ, η n the horzontal and terran-followng σ -coordnate n the vertcal. The metrc coeffcents connectng ( φ, λ to ( ξ, η are defned by λ = φ + ϕ g ( cos ( λ = + g ( cos (, ( ϕ. ( The coeffcent g s the metrc tensor n ξ -drecton and the coeffcent g metrc tensor n η -drecton. These tensors permt the model to transform the user defned boundary ftted grd to a numercal grd employed for spatal dscretzaton utlzed n an Arakawa C Grd. Let (ζ, H be the surface elevaton and bathymetry. D = H +ζ, s the total water depth. The σ - and z- coordnates are connected by z + H σ =, (3 ζ + H whch makes σ = for the ocean surface and σ = for the ocean bottom. The D WQMAP represents a depth-averaged shallow-water system (smlar to Wang et al., 998. Let (U, V be the vertcally averaged velocty components n ( ξ, η drectons. The momentum equatons for (U, V are gven by UD + t ( g ( UD g ( UVD g ( g gd ζ D ρ σ D ρ ] fdv = + dσ R g ρ D ξ σ σ g g V w b + ( τ τ + AD U, ξ ξ h ρ ( [ + + UVD ( UVD g ( V D g ( g VD + [ + + UVD t g g g gd D D ] + = + U w b + ( τ τ + AD V. η η h ρ ζ ρ σ ρ fdv dσ R g ρ D σ η σ The contnuty equaton s represented by (4 (5 3
R g g ( UD g ( VD g ζ + + t =. (6 Here, R s the earth radus; ρ (= 5 kg m -3 s the characterstc densty for the seawater; f s the Corols parameter; g s the acceleraton due to gravty; A h s the w w horzontal eddy vscosty; ( τ, τ are the wnd stress; and ξ b b τ, τ are the bottom stress. As wth any depth-averaged ( ξ η model, t s mplctly assumed that velocty and densty are nearly constant over the water column. However, horzontal densty gradents are treated explctly n the momentum equatons. As we mentoned n Secton that the freshwater flow and surface wnds n the bay are low. The currents n San Dego Bay are predomnately produced by tdes. Thus, the horzontal densty gradent can be neglected n short-term predcton. 4. MODEL IMPLEMENTAION WQMAP for San Dego Bay covers an area of 43 km. Dfferent from Wang et al. (998, the model doman s only for the bay and the boundary fttng orthogonal curvelnear coordnates are used (Fg. 3. The computatonal mesh has 5 5 (,5 grd nodes wth an average horzontal resoluton of 45 m. Model bathymetry s determned from depth soundng data provded by NOAA and supplemented by data from publshed navgaton charts. Recently Navy conducted bathymetry surveys show that the water depths n regons near the bay entrance are sgnfcantly deeper than the water depths shown on the NOAA navgaton chart (Wang et al., 998. The most up-to-date bathymetry data are used n the model (Fg. 4. Water surface elevaton and velocty are set to zero, and temperature and salnty are assgned as the characterstc values for San Dego Bay (6 o C, 34 ppt at all grd ponts. The model s allowed to spn up from quescent ntal condton for one day before any model results are used for analyss. A sx-mnute tme step s chosen for tme step. At ths tme step the CFL condton s satsfed. Besdes, the model parameters are gven as follows: the wnd drag coeffcent (.4, the bottom drag coeffcent (.3, the vertcal vscosty (.5 m s -, the vertcal dffusvty (. m s -, and the horzontal dffusvty (. m s -. η Fg. 3. Model grds. Fg. 4. San Dego Bay bathymetry. 5. TIDAL FORCING Temporally varyng sea surface elevaton (or tdal harmonc consttuents along the open boundary (entrance of San Dego Bay s taken as the model forcng functon. Such data are avalable at the NOAA Center for Operatonal Oceanographc Products and Servces webste. The elevaton data wth sx-mnute nterval are archved from tme on January 4 to 354 on 3 January 4 for San Dego Bay entrance (Staton Identfcaton Number 947,.e., Staton- n Fg.. Ths data set conssts of 7,44 lnes of observed elevaton data ponts and contans hgh-frequency non-tdal sgnals. The fast Fourer Transform s conducted on the data set, and the hgh frequency sgnals that was present wthn the data set from wnds, fronts, etc were fltered out. The new smoothed elevaton data are used as the tdal forcng functon at the entrance of San Dego Bay. 4
6. COMPATIBILITY VERIFICATION Usually, the NOAA tdal table lsts the surface elevaton (6-mn nterval and maxmum ebb and flood current speeds (4 tmes daly. When the surface elevaton data at the open boundary (.e., Staton- n Fg. are used as the tdal forcng to drve the hydrodynamc model (.e., WQMAP, temporally varyng horzontal current velocty wth horzontal resoluton of around 45 m s generated for the whole San Dego Bay. Snce the maxmum ebb and flood current veloctes at the open boundary (.e., Staton- from the NOAA tde table are not used to force the numercal model, t s mportant to check f the model generated current veloctes near Staton- consst wth that n the NOAA tde table. Ths process s called the compatblty verfcaton. If the model generated current veloctes near the open boundary are dfferent from that lsted n the NOAA tde table, the model s dynamcally ncompatble at the open boundary. Otherwse, the model s dynamcally compatble at WQMAP NOAA the open boundary. Let [ v ( t, v ( t ] be the model generated and the NOAA tde table lsted ebb and flood current veloctes. The mean relatve error (MRE between the two WQMAP NOAA v ( t v ( t MRE =, (7 NOAA M v ( t s used to dentfy the dynamcal compatblty at the open boundary. Here, M s the total number of data ponts. The less the MRE, the more dynamcally compatble at the open boundary s. Usually, MRE < % can be used as the ndcaton for the dynamcal compatblty. The NOAA tde table current velocty data used for the compatblty test (at Staton- are n reference to a postve value for the flood drecton of 355 and a negatve value for the ebb drecton of 75. The model-generated horzontal velocty n zonal and lattudnal components (u, v should be transformed nto (u, v n the ebb and flood drectons for comparson wth the NOAA tde table. Fg. 5 shows the model generated (dashed curve and the NOAA tde table lsted (sold curve ebb and flood currents at the open boundary (.e., Staton-. The modeled velocty s hghly n-phase wth the NOAA tde table data. Modeled results are slghtly over predcted n the flood currents and smlarly under predcted n ebb currents. The ntal large error n ampltude and phase s seen at the begnnng of the model run and s prmarly due to the ramp up tme needed for the model to ntalze. The model elevaton forcng was already n place at the nstant of the model run but ntal current speed for the model ntalzaton was zero meters per second. After a perod of one day, the model nearly perfectly matches the phase and ampltude of the verfcaton data. The MRE between the modeled currents and the NOAA tde table data set s 4.%, whch shows the dynamcal compatble at the open boundary. 5 Fg. 5. Comparson between model generated and NOAA tde table lsted maxmum ebb and flood current veloctes at Staton- (compatblty verfcaton. Besdes the compatblty test, the model generated current velocty nsde the bay should be compared wth the observatonal data such as the current meter data, acoustc Doppler current profle (ADCP data, etc. ([6], [7]. Snce the NOAA tde table s the easest envronmental source to obtan for most naval operatons, and snce the purpose of ths study s to show the mportance of the lttoral zone modelng for Navy s operatons, detaled verfcaton s not descrbed here. 7. APPLICATION TO THE MIW EXERCISES Secton descrbes a smple mne warfare operaton was conducted near Staton-4 nsde San Dego Bay on January 4. Wth the maxmum ebb and flood current veloctes at Staton- (entrance of San Dego Bay from the NOAA tde table, the mne countermeasure can operate wthn the threshold for current velocty ( m s - all mornng untl :. The mne huntng would be able to resume agan around 3:45 wth no other restrctons for the rest of the day (Fg.. Fg. 6 shows the tme seres of the predcted ebb and flood current velocty at Staton-4 on the same day ( January 4 usng WQMAP wth the tdal forcng at the open boundary (Staton-, from NOAA tde table. The model predcted current speed at Staton-4 s much slower than the tdal current speed at Staton- lsted n NOAA tde table. Fg. 7 shows that the nstantaneous speeds are all less than the threshold ( m s -. There s no need to nterrupt mne countermeasure and mne huntng operatons. Fg. 6. Comparson between model generated (at Staton-4 and
NOAA tde table lsted (at Staton- maxmum ebb and flood current veloctes. Ths example clearly llustrates the value added ths hydrodynamc program can provde to effectve msson plannng. Fg. 6. Comparson between model generated (at Staton-4 and NOAA tde table lsted (at Staton- maxmum ebb and flood current veloctes. Fg. 7. Msson plannng for mne countermeasure and mne huntng nsde San Dego Bay (Staton-4 usng model (WQMAP generated current veloctes. January, 4: NOAA Tde Table Currents 8. APPLICATION TO THE DIVING OPERATIONS An addtonal and more useful example s to extend the same operaton on the same day to nclude dvng operatons. One opton to mne neutralzaton s to recover the mne usng an explosve ordnance dsposal (EOD team. The scenaro contnues wth the mne countermeasure fndng and dentfyng a mne lke object as an actual mne. The mne was found near a structural pylon of the Coronado Brdge (Staton-4 so detonatng t usng the mne neutralzaton vehcle s not a desrable opton. An EOD team s dspatched to retreve the mne. Ther operatonal threshold for currents s also operatonally senstve so a fcttous lmt of.5 m s - (.e.,.5 knots wll be used for ths example. There are four operatonal wndows for dvng operatons n ths scenaro usng the NOAA tde table data set at Staton- (Fg. 8. They range from 3 to 3, 8 to 9, 53 to 655, and 3 to 33. Agan, compare these wndows to those found usng the modeled currents (Fg. 9 for the same locaton and the operaton pcture changes. The revsed operatonal wndows due to the modeled currents are 3 to 3, 8 to 9, 65 to 73, and 3 to 3. The frst two wndows were reduced slghtly. More crucal changes occurred to the last two wndows. The thrd wndow predcton was moved forward n tme by an hour and twenty mnutes and was reduced n duraton from an hour and twenty-fve mnutes down to just forty mnutes. Such changes n operatng wndows are crucal to successful operatonal plannng to ensure msson success and safety to all personnel nvolved. 6 Fg. 8. Msson plannng for the dvers operatons nsde San Dego Bay (Staton-4 usng NOAA tde table maxmum ebb and flood current veloctes at Staton-. Fg. 9. Msson plannng for dvers operatons nsde San Dego Bay (Staton-4 usng model (WQMAP generated current veloctes. 9. CONCLUSIONS Ocean envronmental condtons affect naval operatons. Each warfare communty has defned
envronmental thresholds and operatng lmts that restrct the executon of any ntended maneuver. Accurate predcton of nstantaneous velocty s essental n the lttoral warfare such as n mne countermeasure, mne huntng, and dvers operatons. The value-aded of usng a hydrodynamc model (.e., WQMAP for the msson plannng of the naval operatons n San Dego Bay s demonstrated n ths study n two aspects. Frst, the msson plannng changes wthout and wth usng the hydrodynamc model. Wthout the hydrodynamc model, the msson plannng for the mne countermeasure, mne huntng, and dvers operatons nsde the bay s conducted usng the maxmum ebb and flood current speeds from the NOAA tde table (only one avalable at the entrance of the bay,.e., Staton-. Wth the hydrodynamc model, the msson plannng s conducted usng the model-generated nstantaneous current velocty data at the exact locaton (nsde the bay where the naval operatons take place. Second, the hydrodynamc model should be suffcently accurate. Thus, thorough model verfcaton should be conducted before t s used for the msson plannng. For a tdal domnated lttoral basn such as San Dego Bay, two knds of verfcaton should be conducted: ( routne, and ( compatblty verfcaton. The routne verfcaton s commonly used. The modeled data s compared wth the observed data (such as ADCP data nsde the bay. The compatblty verfcaton s to check f the model generated current veloctes consst wth that n the NOAA tde table near the open boundary. Ths study shows that WQMAP for San Dego Bay passes the compatblty verfcaton wth the relatve mean error of 4.%. As the warfare communtes contnue to advance n the realm of lttoral warfare, so wll the support for those communtes. There contnues to be an ncreasng demand to better model the much more complex coastal envronment. The current use of the Naval Oceanographc Offce s fleet survey teams can readly provde what s needed to valdate these models for each lttoral grd that may be needed. The prmary msson of the fleet survey team s to obtan bathymetrc data for navgatonal chart development. The same bathymetry can be mported nto a boundary ftted grd of the same regon to help model ths envronment. The fleet survey team can facltate a successful model valdaton for a coastal area of nterest by obtanng elevaton data durng a survey perod to be used as forcng data for the open boundary cells defned wthn the model. Furthermore, the fleet survey team, by addng an ADCP to ther nventory, can obtan routne verfcaton data durng ther tme frame of ther hydrographc survey. Wth proper plannng, ths asset of the Naval Oceanographc Offce can provde the warfare communtes more than just an mproved navgatonal product. The msson of the fleet survey team n conjuncton wth the Naval Oceanographc Offce s modelng and forecastng dvsons wll be able to produce a seres of small, 7 PC-based current predcton models that wll be an nvaluable beneft to future and unpredctable naval operatons. ACKNOWLEDGMENTS Ths work was jontly supported by the Naval Oceanographc Offce (document number: N6364PO3 and Naval Warfare System Command and the Naval Postgraduate School. REFERENCES [] A. E. Armstrong, 4. Predcton of nstantaneous currents n San Dego Bay for naval applcatons. MS Thess n Meteorology and Oceanography, Naval Postgraduate School, (Monterey, 56 pp. [] P.F. Wang, R.T. Cheng, K. Rchter, E.S. Gross, D. Sutton, and J. W. Gartner, 998. Modelng tdal hydrodynamcs of San Dego Bay, Calforna. Journal of the Amercan Water Resources Assocaton, 34 (5, 3-4. [3] T. J. Peelng, 975. A proxmate bologcal survey of San Dego Bay, Calforna. Naval Undersea Center, San Dego, Calforna, Report No. TP389. [4] M. Mun, and M. L. Spauldng, 996. Two-dmensonal boundary ftted crculaton model n sphercal coordnates. Journal of Hydraulc Engneerng, (9, 5-5. [5] M. Mun, and M. L. Spauldng, 997. Three-dmensonal boundary ftted crculaton model. Journal of Hydraulc Engneerng, 3 (, -. [6] P.C. Chu., S.H. Lu, and Y.C. Chen,. Evaluaton of the Prnceton Ocean Model usng the South Chna Sea Monsoon Experment (SCSMEX data. Journal of Atmospherc and Oceanc Technology, 8, 5-539. [7] P.C. Chu, C.W. Fan, and L.L. Ehret, 997: Determnaton of open boundary condtons from nteror observatonal data," Journal of Atmospherc and Oceanc Technology, 4, 73-734